Gait & Posture 23 (2006) 348354

www.elsevier.com/locate/gaitpost

Postural strategy changes with fatigue of the lumbar extensor muscles

Erin L. Wilson a, Michael L. Madigan a,*, Bradley S. Davidson a, Maury A. Nussbaum b
a
Department of Engineering Science and Mechanics, Virginia Polytechnic Institute and State University, Mail Code 0219, Blacksburg, VA 24061, USA
b
Grado Department of Industrial and Systems Engineering, Virginia Polytechnic Institute and State University,
250 Durham Hall (0118), Blacksburg, VA 24061, USA
Received 1 February 2005; accepted 15 April 2005

Abstract

The purpose of this study was to investigate the effect of lumbar extensor fatigue on postural strategy in response to a balance perturbation.
Anteriorly-directed force perturbations were applied to the upper back with a padded pendulum and attempted to challenge the postural
control system without eliciting a stepping response. In three separate sessions, subjects were perturbed both before and after a fatiguing
protocol that induced lumbar extensor fatigue to one of three different fatigue levels. Postural strategy was quantified using center of pressure
position along with joint angles and joint torques for the ankle, knee, hip, and low back joints. Results showed both proactive and reactive
changes in postural strategy. Proactive changes involved a slight anterior lean prior to the perturbation, and reactive changes were consistent
with a shift toward more of a hip strategy with fatigue. In addition, results suggested that subjects classified as moving mostly at the hip prior to
fatigue were more affected by fatigue compared to subjects classified as moving roughly equal amounts at the ankle and hip prior to fatigue.
Increasing fatigue level exaggerated some, but not all, of the changes in postural strategy with fatigue. These findings illustrate that
neuromuscular fatigue can influence postural strategy in response to a balance perturbation.
# 2005 Elsevier B.V. All rights reserved.

Keywords: Postural strategy; Fatigue; Perturbation; Balance

1. Introduction Neuromuscular fatigue is another intrinsic factor that

Movements used to maintain standing balance in the
sagittal plane following a postural perturbation have been
described as a hip strategy, ankle strategy, or a combination
of the two [1,2]. Numerous extrinsic and intrinsic factors
affect, which of these postural strategies are typically
employed. Extrinsic factors include support surface size [3],
perturbation magnitude [4], perturbation direction [5], and
body position prior to the perturbation [6]. These studies
suggest that postural strategy is affected by the biomecha-
nical demands associated with maintaining balance.
Intrinsic factors include fear of falling [7], expectation of
perturbation characteristics [8], quality of somatosensory
and vestibular feedback [9], Parkinsons disease [10], aging
[11], low back pain [12], and respiration [13].
* Corresponding author. Tel.: +1 540 231 1215; fax: +1 540 231 4574.
E-mail address: mlmadigan@vt.edu (M.L. Madigan).
may affect postural strategy, but to our knowledge no studies
have investigated fatigue in this regard. Fatigue has been
shown to increase postural sway during quiet standing [14
20]. Increases in postural sway have, in turn, been associated
with changes in postural strategy [21,22]. For example, Kuo
et al. (1998) administered the Sensory Organization Test to
healthy subjects and observed an increase in postural sway
along with an increase in hip strategy when somatosensory
information was disrupted. Although this study did not focus
on fatigue, it demonstrates potential effects of a change in
postural strategy on postural sway. Similarly, the docu-
mented increase in postural sway during quiet standing with
fatigue may result from a change in postural strategy. If
fatigue induces changes in postural strategy during quiet
standing, it seems logical that similar changes in postural
strategy will occur in response to a postural perturbation.
The purpose of this study was to investigate the effect of
lumbar extensor fatigue on postural strategy in response to a
balance perturbation. The effect of lumbar extensor fatigue

0966-6362/$ see front matter # 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.gaitpost.2005.04.005
 E.L. Wilson et al. / Gait & Posture 23 (2006) 348354
level was also investigated to determine if increasing levels
of fatigue have increasing effects on postural strategy.
Understanding any changes in postural strategy with fatigue
will further our basic understanding of postural control, and
potentially contribute to the development of interventions
aimed at mitigating the effect of fatigue on balance. One
application of such an intervention would be in the
construction industry where fatigue-related loss of balance
may contribute to over 300 deaths/year from falls from
heights [23]. Lumbar extensor fatigue was investigated
because it is common during many occupational lifting tasks
[24], and muscle fatigue in general has been shown to
increase the risk of falling from heights [25].
2. Methods
Twelve physically active males (2022 years of age)
participated in the experiment. Mean (S.D.) participant
height and mass were 173.7(6.4) cm and 70.2(6.6) kg,
respectively. None of the participants reported any history of
low back pain or injury, and all provided informed consent in
accordance with the Virginia Tech Institutional Review
Board before participation.
Participants completed three experimental sessions with
12 weeks between sessions. Each session consisted of four
stages: warm-up, unfatigued balance perturbations, fati-
guing protocol, and a fatigued balance perturbation. For
warm-up, participants completed two cycles of the follow-
ing activities: jogging at 2.0 km/h on a treadmill for 21/
2 min, stretching the lumbar extensors, and performing five
back extensions on a 458 Roman chair (New York Barbell,
Elmira, NY). Following warm-up, the unfatigued isometric
maximum voluntary contraction (MVC) moment of the
lumbar extensors was measured at the level of L3.
Participants were positioned in the Roman chair, attached
to a load cell (Cooper Instruments and Systems, Warrenton,
VA) at the mid-sternum via a modified construction harness,
and instructed to extend at the back as hard as possible. In
this position the lumbar flexion angle was approximately
458. Participants performed three isometric MVCs separated
by 1 min of rest; the largest of the three was recorded as the
unfatigued MVC value.
Three unfatigued balance perturbations were performed
following the warm-up. Prior to the perturbation, partici-
pants were instructed to stand quietly on a force platform
without shoes, with their eyes closed, feet together, and arms
resting at their side (Fig. 1). Participants were also told that if
they stepped to regain their balance, another collection
would be taken. (Two perturbations were repeated due to
stepping.) Perturbations consisted of an anteriorly-directed
force applied in the mid-sagittal plane at the level of the
inferior margin of the scapulae, and were applied using a
padded pendulum that was pulled a fixed distance away from
the subject and released. These perturbations were designed
to mimic a bump to the upper back and challenge the
349
Fig. 1. Experimental set-up for applying postural perturbations. Subjects
were perturbed by releasing the pendulum after pulling it back a fixed
distance from the subject.
postural control system without eliciting a stepping
response. Using this method, the perturbation peak force
was 177(21) N, and the perturbation impulse was
13.2(1.3) N s. Perturbations were applied following a
random delay (015 s) from the beginning of each data
collection to reduce participant anticipation.
In each of the three experimental sessions, the lumbar
extensor muscles were fatigued to either 86, 73, or 60% of
their unfatigued MVC. A balanced Latin square was used to
determine the order of presentation of the three fatigue
levels. The fatiguing protocol consisted of multiple sets of
back extensions and intermittent MVCs performed on the
Roman chair (see [15] for more details). Investigators
attempted to fatigue participants such that their MVC
moment decreased linearly over the duration of the fatiguing
protocol and achieved the desired fatigue level over 14 min.
The fatigue levels actually achieved were 82(5), 70(5), and
62(6)% MVC, which were significantly different from each
other ( p < 0.001). Prior to performing the back extensions,
participants were instructed to keep their feet flat on the base
of the Roman chair and move through approximately 458
range of motion from 08 lumbar flexion to maximum flexion.
Participants performed one set of back extensions every
minute of the fatiguing protocol and an isometric back
extension MVC every 2 min. MVCs were performed 10 s
after completing the prior set of back extensions. A digital
metronome set at 30 beats/min was used to ensure all
participants performed the extensions at a similar rate. The
number of repetitions in each set was systematically
adjusted based upon a comparison of each measured
MVC moment to the target moment [15]. If participants
were not at or below the desired fatigue level at the end of the
fatiguing protocol, 2 min were added with additional
repetitions according to the same method. This process
350 E.L. Wilson et al. / Gait & Posture 23 (2006) 348354
was continued until the MVC was below the desired fatigue
level.
Within 20 s after completing the fatiguing protocol, one
fatigued balance perturbation was performed in an identical
manner as the unfatigued perturbations. Only one fatigued
perturbation was performed to minimize recovery from
fatigue.
During each trial, the perturbation force was sampled at
1000 Hz from a load cell mounted inline with the pendulum
and subsequently low-pass filtered at 25 Hz (fourth order
zero-phase-lag Butterworth filter). Triaxial ground reaction
forces and moments were low-pass filtered with a 500 Hz
anti-aliasing hardware filter, amplified, and sampled at
1000 Hz with a Bertec K20102 type 9090-15 force platform
(Bertec Corp., Columbus, OH). These data were subse-
quently low-pass filtered at 10 Hz (fourth order zero-phase-
lag Butterworth filter) and transformed to obtain ground
reaction forces (GRF) and center of pressure (COP) data
[26]. Body position was sampled at 50 Hz using a Vicon 460
motion analysis system (Vicon Motion Systems Inc., Lake
Forest, CA). Markers were placed on both sides of the body
at the acromion, iliac crest, greater trochanter, lateral
femoral epicondyle, lateral malleolus, calcaneus, and head
of the fifth metatarsal. Marker position data was low pass
filtered at 7 Hz (fourth order zero-phase-lag Butterworth
filter) and averaged across both sides of the body to create a
sagittal plane representation of the body with five segments
including the feet, shanks, thighs, pelvis, and trunk (defined
by the iliac crest and shoulder). Using these five segments,
joint angles were calculated for the ankle, knee, hip, and
low back joints (the low back joint is defined as the angle
between the trunk and pelvis segments). Joint torques were
estimated using an inverse dynamics analysis with a model
of the body consisting of five rigid segments connected by
frictionless pin joints. The mass and inertial characteristics
of the segments were defined using an anthropometric model
[27].
Three dependent measures were calculated to quantify
preparatory activity prior to the perturbation including initial
joint angles, initial joint torques, and initial COP position
relative to the ankle. These initial values were calculated as
the mean over the time interval from 540 to 40 ms prior to
the perturbation. Four dependent measures were calculated
to quantify postural strategy during recovery from perturba-
tions including peak joint angles, time to peak joint angles,
peak joint torques, and time to peak joint torques. Because
initial joint torques prior to the perturbation were affected by
fatigue (see Section 3), peak joint torques relative to their
initial values were used rather than absolute torques. Only
data from the third unfatigued balance perturbation was used
in the analysis because of possible learning effects within the
three unfatigued perturbations [28].
To determine the effect of fatigue and fatigue level on
postural strategy, two separate two-way repeated measures
multivariate analysis of variance (MANOVA) were used.
The first MANOVA used dependent measures quantifying
preparatory activity prior to the perturbation, and the second
used dependent measures quantifying strategy during
recovery from the perturbation. Data from all four joints
were included in each MANOVA with the independent
variables being fatigue (two levels: unfatigued and fatigued)
and fatigue level (three levels: 86% MVC, 73% MVC, and
60% MVC). A significant main effect was followed by a
two-way repeated measures analysis of variance (ANOVA)
for each dependent variable. Correlation analysis was
performed to determine the strength of any relationship
between measures quantifying preparatory activity prior to
the perturbation and measures quantifying strategy during
recovery from the perturbation. A significance level of
p  0.05 was used for all tests.
To determine if perturbations were consistent across
fatigue states, a two-way repeated measures MANOVA was
also used to determine if perturbation peak force and
impulse were affected by fatigue and fatigue level.
Perturbation peak force decreased 4.3(6.4)% with fatigue
( p = 0.013), and there was no effect of fatigue level or
fatigue  fatigue level interaction for either parameter.
3. Results
Although a fair amount of inter-subject variability was
observed, a general pattern of joint angles and torques
emerged across all subjects during unfatigued trials (Fig. 2).
Prior to the perturbation, initial joint angles were
approximately zero (i.e. anatomical position). Initial ankle
and knee torques were plantar flexor and flexor dominant,
respectively, while the initial hip and low back torques were
more variable. Following the perturbation, the initial
response was a flexor torque at the knee and extensor
torques at the hip and low back. These torques were followed
1030 ms later by a plantar flexor torque at the ankle and
flexion at the knee, hip, and low back. Ankle plantar flexion
began 2080 ms after the ankle torque. The entire recovery
duration was approximately 2 s. The timing of joint angles
and torques were consistent within the unfatigued trials and
were not affected by fatigue, fatigue level, or their
interaction.
Fatigued recoveries exhibited the same general pattern of
joint angles and torques as in the unfatigued recoveries, but
some specific differences existed. Prior to the perturbation,
initial joint angles did not change with fatigue ( p = 0.441),
but the COP position was 1.1(1.7) cm more anterior from the
ankle compared to unfatigued trials ( p = 0.012). The
anterior displacement of the COP during static stance
suggests increased plantar flexor activity and necessitates a
forward lean (the magnitude of which was evidently too
small to be detected in joint angle measurements). As a
result of this slight forward lean when fatigued, initial joint
torques at all four joints changed with fatigue; offsets were
in the direction of the maximum torque for each joint. More
specifically, initial ankle torque shifted 7.1(1.1) N m in the
 E.L. Wilson et al. / Gait & Posture 23 (2006) 348354
Fig. 2. Representative joint angles and torques during recovery from an
unfatigued postural perturbation. Positive values indicate extension move-
ment (relative to anatomical position) and extensor torque. The perturbation
time is indicated with a dotted line. The plots show the general patterns
observed from the subjects.
plantar flexor direction ( p = 0.006), initial knee torque
shifted 4.8(3.8) N m in the flexion direction ( p = 0.010),
initial hip torque shifted 6.5(1.7) N m in the extensor
direction ( p < 0.001), and initial low back torque shifted
6.1(1.4) N m in the extensor direction ( p = 0.001). There
were no fatigue level effects or fatigue  fatigue level
interactions for initial COP position, initial joint angles, or
initial joint torques.
Both peak joint angles and torques after the perturbation
were affected by fatigue (Fig. 3) and fatigue level. Fatigue
increased peak ankle plantar flexion 2.5(0.5)8 ( p = 0.002),
did not affect peak knee ( p = 0.170) or hip ( p = 0.062)
angles, and increased peak low back flexion 6.3(1.6)8
( p = 0.021). In addition, fatigue decreased peak ankle
351
plantar flexion torque 3.2(2.1) N m ( p = 0.002), did not
affect peak knee torque ( p = 0.310), increased peak hip
extensor torque 10.2(1.2) N m ( p = 0.044), and increased
peak low back extension torque 9.4(0.8) N m ( p = 0.013).
Fatigue level also had an effect on several dependent
measures. Increasing fatigue level from 86% MVC to 60%
MVC increased peak low back flexion 4.6(7.0)8 ( p = 0.049),
decreased peak ankle plantar flexor torque 5.8(6.4) N m
( p = 0.030), decreased peak knee flexor torque 6.8(7.0) N m
( p = 0.005), and decreased hip extensor torque 3.8(8.3) N m
( p = 0.034). Fatigue level had no effect on peak low back
torque ( p = 0.376), and there were no fatigue  fatigue level
interactions for peak joint angles or torques.
Correlations between initial COP position and peak joint
angles/torques revealed minimal covariance between these
variables except for one peak torque value. Peak ankle
torque exhibited a small but statistically significant
correlation with initial COP position (r2 = 0.19; p < 0.05).
The coefficient of determination for all other correlations
were r2 = 0.01(0.01), and none were statistically significant.
Although all subjects exhibited the same general
recovery patterns that are best described as a mixed strategy,
subjects could be classified into two groups based on the
amount of movement at the hip and ankle during unfatigued
recoveries (Fig. 4). Through visual inspection of the
unfatigued data, eight of the twelve subjects were found
to move predominantly at the hip and low back and were
classified as using more of a hip strategy. The remaining four
subjects showed similar amounts of movement at the ankle,
hip, and low back, and were classified as using a true mixed
strategy. The effects of fatigue on postural strategy were
more pronounced in the hip strategy group. For the hip
strategy group, fatigue increased peak ankle plantar flexion
2.4(3.4)8, increased hip flexion 5.2(8.0)8, and increased low
back flexion 8.4(11.0)8. For the mixed strategy group,
fatigue increased peak ankle plantar flexion 2.9(2.8)8,
increased hip flexion 1.9(2.1)8, and increased low back
flexion 3.5(3.9)8, respectively.
4. Discussion
Neuromuscular fatigue has been shown to increase
postural sway during quiet standing [1420]. The increase in
sway may, in part, be the result of a change in postural
control strategy. The purpose of the present study was to
investigate the effects of lumbar extensor fatigue and fatigue
level on postural strategy in response to a balance
perturbation. Our results showed both proactive and reactive
changes in postural strategy with fatigue. For the purposes of
this report, proactive changes are operationally defined to be
changes occurring prior to the perturbation, and reactive
changes are the changes occurring after the perturbation.
Proactive changes involved adopting a slight forward lean
prior to the perturbation as evidenced by an anterior shift in
the COP, while reactive changes indicate a shift toward hip
352 E.L. Wilson et al. / Gait & Posture 23 (2006) 348354

Fig. 3. Representative joint angle and joint torque data during unfatigued and fatigued postural recoveries. Positive values indicate extension movement
(relative to anatomical position) and extensor torque. Joint angle plots show increases in peak ankle plantar flexion and peak low back flexion with fatigue. Joint
torque plots show initial torque offsets and increases in ankle, hip, and low back extensor torques with fatigue. Figures in the middle illustrate differences in
unfatigued and fatigue recoveries at the instant of maximum hip flexion.

strategy. Increasing fatigue level exaggerated some, but not
all, of these changes. Interestingly, the changes were not
limited to the fatigue location.
The proactive change in postural strategy involved a
1.1(1.7) cm anterior displacement of the COP prior to the
perturbation. Although a post hoc analysis of the angle
between the ankle and whole-body center-of-mass (COM)
prior to the perturbation did not show a significant increase
Fig. 4. Differences in peak joint angles and torques between the hip and
mixed group. Both line plots show a greater affect of fatigue on the hip
group as compared to the mixed group. The stick figures show the general
patterns of movement for both groups. The dotted figure is the initial
(anatomical) position while the solid figure is the maximum displacement
from the initial position following the perturbation. Standard deviation bars
were omitted for clarity, but these data are available from the authors.
with fatigue, mechanics necessitates at least a slight forward
lean due to the COP shift.
In both unfatigued and fatigued recoveries, subjects
generally used a mixed strategy; however, reactive changes
with fatigue showed a shift toward more of a hip strategy.
Changes in peak torques indicated a clear shift toward a hip
strategy, as evidenced by a 3.2 N m decrease in plantar flexor
torque and a 10.2 N m increase in hip extensor torque. The
2.58 increase in peak ankle plantar flexion is consistent with
this shift, since the hip strategy involves simultaneous hip
flexion and ankle plantar flexion. Although the increase in
peak hip flexion was only a trend ( p = 0.062), peak low back
flexion increased 6.38, which provides a similar benefit to
recovery as hip flexion by attempting to reduce the body
moment of inertia about the ankle [4]. Taken together, we
feel that these results indicate a shift toward more of a hip
strategy with lumbar extensor fatigue. It was also interesting
to note that peak low back extension torque increased
9.4 N m despite subjects being fatigued.
According to past studies [5], the change in reactive
postural strategy could be partially attributed to the forward
lean. However, three arguments suggest otherwise. First, low
correlations between initial COP position and peak joint
angles/torques indicate a weak relationship between these
measures. Second, a follow-up analysis was performed by
repeating our statistical analysis on a subset of participants
whose change in initial COP with fatigue was less than the
overall mean of 1.1 cm (seven participants). This subset of
participants showed no change in initial COP position with
fatigue ( p = 0.286), and significant changes in reactive
measures with fatigue that were similar to the overall
 E.L. Wilson et al. / Gait & Posture 23 (2006) 348354
analysis: peak ankle angle increased ( p = 0.046), peak hip
extensor torque increased ( p = 0.031), and peak low back
extensor torque increased ( p = 0.012). Third, comparing
lean angles in the present study with those of another study
suggest that the leans here were too small to elicit a change
in recovery strategy. Horak and Moore [6] reported that a hip
strategy is favored when pre-leaning in the same direction as
an induced sway. However, this shift in strategy was only
found when subjects assumed a maximum forward lean (an
average 6.098 from normal standing posture). No change in
strategy was found when subjects assumed a half-maximum
forward lean (an average 3.018). The average forward lean in
our study was 0.258, which is much smaller than either of the
leans reported. For these reasons, we do not attribute the
reactive changes in postural strategy to the forward lean
prior to the perturbation.
Reasons why a shift toward more of a hip strategy and an
increase in forward lean were observed with lumbar extensor
fatigue remain to be explained. In regards to the forward
lean, we feel it is unlikely that lumbar extensor fatigue
directly elicits an increase in plantar flexor activity without
any other biomechanical or physiological explanation. One
reason for the lean, independent of fatigue, may be that
subjects unintentionally leaned forward with repeated
exposure to the balance perturbation. Such a habituation
has been reported [8]. This habituation, however, seems to
occur mostly in the first few perturbations [28], and we
allowed 3 unfatigued perturbations to allow for habituation
prior to fatiguing subjects. Also, there was no difference in
COP initial position between the second and third
unfatigued perturbations ( p = 0.907), suggesting that habi-
tuation had already occurred. Another potential reason is
related to motor control. Lumbar extensor fatigue has been
associated with an increase in antagonist muscle activity in
the rectus abdominus muscle (RA) [29]. We analyzed
muscle activity data from five subjects that was collected for
a separate study and found baseline RA activity increased an
average of 54% with lumbar extensor fatigue prior to the
perturbation ( p = 0.033). Forward leaning has also been
associated with an increase in RA activity [6]. The increase
in RA activity prior to the perturbation may be a part of a
subtle adaptation in postural control that displaced the COP
forward. Interestingly, Runge et al. [4] reported higher hip
torques (i.e. more of a hip strategy) during perturbation
recoveries that involved elevated RA activity (although their
perturbation method involved a moving platform rather than
a trunk impulse). Thus, an increase in RA activity from
lumbar extensor fatigue has the potential to contribute to
both the forward lean prior to the perturbation and the shift
toward more of a hip strategy in response to the perturbation.
The last potential explanation is attractive because it links
the effects of lumbar extensor fatigue with both the proactive
and reactive changes in postural strategy observed.
It seems reasonable that increasing levels of fatigue
would lead to larger influences on postural strategy. Such a
dose-response relationship implies that increasing fatigue
353
level would exaggerate the effect of fatigue. This was the
case with peak low back flexion, peak ankle plantar flexor
torque, and peak knee flexor torque. The decrease in ankle
movement and increase in low back movement is consistent
with more of a hip strategy. This was not the case, however,
with peak hip extensor torque. Fatigue increased peak hip
extensor torque, but increasing fatigue level from 86% MVC
to 60% MVC decreased peak hip extensor torque an average
of 3.85 N m. In other words, increasing fatigue level
decreased the effect of fatigue on the hip. This may reflect
a continuous adaptation of postural strategy with increasing
level of fatigue. It is interesting to note that the only joint that
did not exhibit a change in peak torque with increasing level
of fatigue was the joint that was fatigued.
It appears that even in the absence of any influencing
factors, differences in preferred postural strategies exist
between individuals. Subtle differences in peak joint angles
during unfatigued perturbations allowed subjects to be
categorized into two strategy groups (Fig. 4). Horak and
Nashner [30] similarly reported that some subjects
employed pure ankle or hip strategies while others used a
mixed strategy even after 2040 trials. In the present study,
there was a tendency for subjects using more of a hip
strategy when unfatigued to be more heavily influenced by
fatigue compared to subjects using a mixed strategy when
unfatigued. These data suggest that an individuals preferred
postural strategy may dictate how much they are affected by
certain influencing factors. Postural strategy may be related
to postural performance and the ability to prevent a fall
after a perturbation, but future work is needed to establish
this relationship.
Several methodological issues warrant discussion. First,
one may question whether the measured changes in postural
strategy were due to fatigue or the 4.3% decrease in
perturbation amplitude. Runge et al. [4] reported greater
shifts toward hip strategy with increasing perturbation
amplitude (i.e. increases in platform velocity). Our
perturbation amplitude decreased with fatigue by an average
of 8 N (decrease in perturbation force). Because of this, we
feel that if perturbation amplitude did not decrease with
fatigue, the measured shift toward hip strategy would have
been larger. The reason for this change in perturbation
amplitude with fatigue is beyond the scope of this study, but
may be related to a decrease in trunk stiffness with lumbar
extensor fatigue. Second, one may question whether the
measured changes in reactive strategy were due to changes
in initial COP position. While we cannot rule this out, we
feel the low correlations between the initial COP position
and reactive measures, along with the result from our follow-
up analysis (individuals who had no change in initial COP
position still exhibited a shift toward hip strategy with
fatigue), provide strong evidence supporting the fact that
almost all of the reactive changes were not due to changes in
initial COP position. Third, all learning affects were
assumed to have happened during the unfatigued balance
perturbations. Fourth, we acknowledge that our relatively
354 E.L. Wilson et al. / Gait & Posture 23 (2006) 348354
low marker sampling rate (50 Hz) may have limited our
ability to identify subtle changes in postural strategy
temporal characteristics.
In conclusion, lumbar extensor fatigue elicited a change
in postural strategy in response to a perturbation. Both
proactive and reactive postural compensations were identi-
fied and show a general shift toward hip strategy with
fatigue. These findings illustrate that neuromuscular fatigue
can influence postural strategy similar to previously reported
intrinsic factors. Future studies are needed to determine how
these strategies affect the ability to maintain balance
following a postural perturbation.
Acknowledgements
This work was supported by grant # J-689 from the
Jeffress Memorial Trust, Richmond, VA (to MLM) and grant
# R01 OH007882 from the Centers for Disease Control and
Prevention (to MAN). Its contents are solely the respon-
sibility of the authors and do not necessarily represent the
official views of the sponsor.
References
[1] Horak FB, Nashner LM. Central programming of postural movements:
adaptation to altered support-surface configurations. J Neurophysiol
1986;55(6):136981.
[2] Kuo AD, Zajac FE. Human standing posture: multi-joint movement
strategies based on biomechanical constraints. Prog Brain Res
1993;97:34958.
[3] Horak FB, Nashner LM. Central programming of postural movements:
adaptation to altered support-surface configurations. J Neurophysiol
1986;55(6):136981.
[4] Runge CF, Shupert CL, Horak FB, Zajac FE. Ankle and hip postural
strategies defined by joint torques. Gait Posture 1999;10(2):16170.
[5] Moore SP, Rushmer DS, Windus SL, Nashner LM. Human automatic
postural responses: responses to horizontal perturbations of stance in
multiple directions. Exp Brain Res 1988;73(3):64858.
[6] Horak FB, Moore SP. The effect of prior leaning on human postural
responses. Gait Posture 1993;1:20310.
[7] Brown LA, Frank JS. Postural compensations to the potential con-
sequences of stability: kinematics. Gait Posture 1997;6:8997.
[8] Maki BE, Whitelaw RS. Influence of expectation and arousal on
center-of-pressure responses to transient postural perturbations. J
Vestib Res 1993;3(1):2539.
[9] Horak FB, Nashner LM, Diener HC. Postural strategies associated
with somatosensory and vestibular loss. Exp Brain Res 1990;82(1):
16777.
[10] Horak FB, Nutt JG, Nashner LM. Postural Inflexibility in Parkinsonian
Subjects. J Neurol Sci 1992;111(1):4658.
[11] Wu G. Age-related differences in body segmental movement during
perturbed stance in humans. Clin Biomech (Bristol Avon)
1998;13:3007.
[12] Hamaoui A, Do MC, Bouisset S. Postural sway increase in low back
pain subjects is not related to reduced spine range of motion. Neurosci
Lett 2004;357(2):1358.
[13] Hamaoui A, Do M, Poupard L, Bouisset S. Does respiration perturb
body balance more in chronic low back pain subjects than in healthy
subjects? Avon Clin Biomech Bristol 2002;17(7):54850.
[14] Yaggie JA, McGregor SJ. Effects of isokinetic ankle fatigue on the
maintenance of balance and postural limits. Arch Phys Med Rehabil
2002;83(2):2248.
[15] Davidson BS, Madigan ML, Nussbaum MA. Effects of lumbar
extensor fatigue and fatigue rate on postural sway. Eur J Appl Physiol
Occup Physiol 2004;93:1839.
[16] Nussbaum MA. Postural stability is compromised by fatiguing over-
head work. Amer Ind Hyg Assoc J 2003;64:5661.
[17] Schieppati M, Nardone A, Schmid M. Neck muscle fatigue affects
postural control in man. Neuroscience 2003;121(2):27785.
[18] Sparto PJ, Parnianpour M, Reinsel TE, Simon S. The effect of fatigue
on multijoint kinematics, coordination, and postural stability during a
repetitive lifting test. J Orthop Sports Phys Ther 1997;25(1):312.
[19] Nardone A, Tarantola J, Giordano A, Schieppati M. Fatigue effects on
body balance. Electroencephalogr Clin Neurophysiol
1997;105(4):30920.
[20] Lepers R, Bigard AX, Diard JP, Gouteyron JF, Guezennec CY. Posture
control after prolonged exercise. Eur J Appl Physiol Occup Physiol
1997;76(1):5561.
[21] Kuo AD, Speers RA, Peterka RJ, Horak FB. Effect of altered sensory
conditions on multivariate descriptors of human postural sway. Exp
Brain Res 1998;122(2):18595.
[22] Speers RA, Kuo AD, Horak FB. Contributions of altered sensation and
feedback responses to changes in coordination of postural control due
to aging. Gait Posture 2002;16(1):2030.
[23] Hinze J. Construction safety. In: Upper Saddle River. NJ: Prentice-
Hall, 1997.
[24] Dolan P, Adams MA. Repetitive lifting tasks fatigue the back muscles
and increase the bending moment acting on the lumbar spine. J
Biomech 1998;31(8):71321.
[25] Hsiao H, Simeonov P. Preventing falls from roofs: a critical review.
Ergonomics 2001;44(5):53761.
[26] Winter DA. Biomechanics and motor control of human movement,
2nd ed., New York: Wiley-Interscience, 1990.
[27] de Leva P. Adjustments to zatsiorsky-seluyanovs segment inertia
parameters. J Biomech 1996;29(9):122330.
[28] Wu G. Examination of trial-independent characteristics of body
kinematics in response to similar postural perturbations. Gait Posture
1998;7(2):1106.
[29] Granata KP, Orishimo KF, Sanford AH. Trunk muscle coactivation
in preparation for sudden load. J Electromyogr Kinesiol 2001;11(4):
24754.
[30] Horak FB, Nashner LM. Central programming of postural movements:
adaptation to altered support-surface configurations. J Neurophysiol
1986;55(6):136981.